<![CDATA[This study assesses the effectiveness of project-based digital learning in which students design a prototype of an Arduino-based intelligent watering and capacitive soil moisture sensor by following the ADDIE flow. The design used was ex post facto descriptive-correlational in one class totaling 117 students, conducted in Indonesia, where traditional farming methods dominate. Achievement data was taken from six components of the assessment (participatory activities, project scores, assignments, quizzes, middle exams, and final exams). Descriptive statistics are visualized through histograms and box-plots; Pearson correlation tests and multiple linear regression were performed to predict the contribution of each component to the final exam, while the technical validity of the prototype was evaluated in ten laboratory cycles using p-charts and Kaplan Meier analysis. The results showed that the value of clustered projects was in the high category (M = 86.8; SD = 7.1) and significantly correlated with the final exam (r = 0.54; β = 0.22; p < 0.01), explaining ± 22 % additional cognitive variance after the formative test was controlled. Participatory activities had a moderate correlation with the final exam (r ≈ 0.50), signaling a transfer of problem-solving skills acquired during prototype collaboration. Tool reliability tests showed 70% of prototypes were in stable function within the first ten cycles; All points were within control limits although the three initial failures hinted at the need for advanced sensor calibration. Learning constraints such as low quiz and assignment scores and wiring/coding difficulties were addressed through micro-deadlines, quick feedback, pair-programming clinics, and prerequisite videos. These findings confirm that ADDIE-based IoT projects can simultaneously improve students' knowledge, practical skills, and problem-solving disposition, provided they are accompanied by pedagogical supports and ongoing technical validation.]]>
Copyrights © 2025
